People have built a reputation on getting in camera effects like that :)

We do have a solution. Use the new Creamsource uber powerful LED light. It'll accept black burst from a camera and sync strobe to shutter. Also does lightning strikes synced to shutter. On the downside the Creamsource is one expensive light but man is it bright, think 575W HMI bright.

On the other hand the fastest cameras made are CMOS. The problem seems to be not so much in the CMOS technology but how it's used.

Bob.

WOW, that looks really bad.

Next generation Cmos will hopefully fix this bull$#@&. It's a big problem that the fanbois can not wish away. This is the worst I've ever seen. It's unacceptable.

The next crop of cams will brag how they don't have this problem. Feel sorry for those with these ex cams. Glad it is not me. I don't have the time or money to waste on this. A LED light will not fix this and the guests at the wedding wouldn't appreciate 500 watts either. Get a clue. It is the technology going backwards.

Well, like with any tool you need to know its limitations.

I'm pretty blown away what one CMOS delivers against 3 CCD, but... The CMOS isn't perfect and if anything this job reminded me (again thankfully, it wasn't my ass on the line) that you can't go on autopilot and have to consider the environment before settling on the gear.

"A LED light will not fix this and the guests at the wedding wouldn't appreciate 500 watts either. Get a clue. It is the technology going backwards. "

You reckon huh?

Hollywood for years have been using lights synced to shutters.
Now we have an almost affordable ($8,000) LED lights capable of the same thing. As for the guests how friggin bright do you think the average strobe light is???

A pretty weak 10J strobe at 1/1000 sec flash duration gives a 10,000W pulse of light. I'd be more concerned that a 400W LED would have the opposite problem, not be bright enough.

Bob.

Bob, you are splitting the hair on issue that's not related. It was in a dance club where we couldn't control the lighting thus your solutions wouldn't work in the first place.

stupid question, but if it's a dance club & it was known the lighting wouldn't be able to change, wasn't there a quick test a day or two ahead of time to see what would happen & if it would be possible to fix it? With 5 cameras this doesn't sound like a last minute, inexperienced setup.

But it (should?) be well known that you either need to light for live or for cameras, you can't really do both.

Patryk, I have to say, I'm intrigued by the artistic possibilities of that look.
I wonder what it looks like sped up.

Patyrk,
"splitting hairs" not really, "flogging a dead horse" more likely :)

I've had to deal with footage shot with CCD cameras and strobe lights. While not as big a mess as that, still a mess. Shooting at 50i you get only one field lit. Not a problem with still camera flashes unless there's a lot of them but with a dance floor strobe running at 10Hz it is quite a mess and it get worse after further compression. I had one stage sequence that used UV strobes on dancers in white satin, impossible to wrangle and encode.

The Happy Friar has it nailed. Any DP worth his salt should have been testing this before the shoot. Not that I should talk, I've been caught out by lighting in a nightclub that looked OK to the eye but was "interesting" when the EX1 tried to shoot it. Thankfully it was actually a cool kind of effect that only hit the background.

To be honest I suspect no matter what camera and lighting was used there'd be problems. The only solution would have been to walk from the job or tell the client the lighting had to be managed for the camera. Easy enough to say in hindsight.

Bob.

Just got back from a 5 camera shoot (Sony EX3) where i was one of the camera ops. I'm so thankful that i wasn't the DP because i would be kicking myself in the ass for picking CMOS based camera on a shoot in a dance club filled with strobe lights as it produced visible dark lines in the footage.

Your terminology was a bit hard to understand .. until I watched the video and Googled rolling shutter CMOS and found the real reason is that the entire frame is not exposed at the same time. It's very similar to a high speed focal plane shutter on a film camera .. where, effectively, just a "slit" is open to the film at any one time as the exposure is recorded. (this can also result in distortion of fast moving subjects across the frame .. as the subject moves during the exposure and is recorded on different parts of the sensor/film)

In still film cameras this effect can be negated by slowing down the shutter to a point where the entire shutter is open when the strobe fires (and of course the strobe is synced to the shutter)

I have no experience with CMOS video so I don't know if this would help (slowing down the shutter) or not. In a dance club you would not be able to "sync" the shutter to the strobe either.

Looks like the effect you got when filming CRT monitors, in ye old days

I am with Farss; people get paid big money to get an effect like that :-)

I suspect that there may be a shutter speed that will reduce it, but it would need to be worked with in advance to see. I agree. Glad I wasn't the DP.

I think it's a pretty cool effect and way more interesting than a simple strobe. I bet you'll have a lot of people saying, "Wow! How did you do that?!?!" rather than, "Wow! You really screwed that up!" I'd call it "art" and charge extra.

love the questions and observations! And Farss usually hits the point. I think there are some workable(maybe weak) solutions in camera and with duration time of strobe units(on some) vs rotating mirror speed( reflective surfaces). With five cameras, with five knowledgeable camera guys, I really amazed nobody review any clips at the start? Or that they all were clueless about lighting, strobes and flashes?

It looks like an engineered effect more that a mistake, but you can always clip the the strobe frames. use PS to adjust the curve of the W.L. brightness .-) then discover the difference, then use a F.F. analysis to subtract or dampen those wavelengths where they exist, smoothing the difference between frames, Hollywood style! Or generate a longer fading strobe and integrate the effect over the flashes!

To go into a shoot without understanding the technical issues involved, in this case strobes and CMOS, is not something to boast about. This is an old issue well understood by stills photographers. I will comment that without the effect the clip would be uninspiring.

As stated before I was not the DP on that 5 cam shoot, but even if i was i most likely would miss the issue. The DP on this shoot was very experienced and i think as it often happens to us he went on autopilot and... Well, shit happened. I'm just passing along the info and the warning that even on seemingly easy jobs there is a way to majorly mess up.

Sorry Apit, that can NOT be fixed.

Dgates, you may be right about post, especially in vegas. But the link was about the clip shot on an EX1 with shutter setting set to off, 1080 30p. The problem is the shutter is "off" and a very fast, short duration strobe used it appears. EX1's multi column.has an unique and very fast data captured but a closer frame by frame study would be more useful because it seems only a pattern of light(flash) moves thru the frame, no people with distorted bodies, walls..etc....

I was suggesting a post process to soften the strobe effect directly, frame by frame, but it is more complex, more of a combustion plugin app at best---- frame by frame. There appears to be enough "data" in the flash region to work with, but it would be tough and costly.

I was attending my niece wedding last Saturday at MSU and noticed the JD had a series of strobes that bounced light of a rotating cylinder( not a ball) with rectangular mirrors with different grading on them. It appeared to created an unique effect, . Watching Pat's linked clip, reminded me of this JD's visual strobe effect.

No video camera, regardless of the sensor technology, is going to accurately capture a flashing strobelight illuminated scene.

John_Cline, there is a world of difference between not seeing every flash (with CCD camera) and seeing the flash captured only in part f the frame (CMOS sensor).

"shutter setting set to off," this is the cause. the sensor columns not currently not being read by the A/D circuit are still receiving light. Whether its an electronic or physical shutter, ccd and film must received no more light for an accurate capture. With film, it would be like a extremely slow mech shutter closure, like a garage door closing.

Yes, there's a difference, but neither is an accurate representation of what the human eye would see. In the case of this particular video, it's a bunch of people dancing, some of them in animal suits and other costumes. Under those circumstances anything goes and like I said, I find the rolling shutter effect more visually interesting than just the stuttering mess of your typical strobe/dance footage.

Maybe the DP did know what was going to happen and that's exactly what he wanted. You know it isn't "right", but maybe the client doesn't. I wouldn't point it out as a "mistake", play it for the client and see if they have the same reaction that I did, they might love it and think you guys are geniuses. If they hate it, then you can apologize. If you apologize before they see it, then there's no way they're going to like it.

I did try repairing it in Vegas using motion blur. Gets pretty funky when the camera moves quickly. There's very significant over exposure from the strobe. The black bars turn into a rolling bar. The overexposure becomes solarisation in parts. Could be some interesting effects to be had if you could control it.

Shutter speed is irrelevant. The EX has a fixed readout time of 1/60", that remains the same no matter the shutter speed or if the shutter is off.

Realistically you cannot capture a strobe light. Even film would lack the latitude and you'd need a high fps to capture it. Strobe lights I've used have a very short duration, only limited by the internal impedance of the capacitors.

Bob.